Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

LR scheduler for SGD in NeuroChem trainer #282

Merged
merged 1 commit into from
Aug 7, 2019
Merged

LR scheduler for SGD in NeuroChem trainer #282

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 18 additions & 11 deletions torchani/neurochem/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -597,20 +597,25 @@ def run(self):
AdamW_optim = AdamW(self.weights, lr=self.init_lr)
SGD_optim = torch.optim.SGD(self.biases, lr=self.init_lr)

scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
AdamW_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
AdamW_optim,
factor=0.5,
patience=100,
threshold=0)
SGD_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
SGD_optim,
factor=0.5,
patience=100,
threshold=0)

while True:
rmse = self.evaluate(self.validation_set)
learning_rate = AdamW_optim.param_groups[0]['lr']
if learning_rate < self.min_lr or scheduler.last_epoch > self.nmax:
if learning_rate < self.min_lr or AdamW_scheduler.last_epoch > self.nmax:
break

# checkpoint
if scheduler.is_better(rmse, scheduler.best):
if AdamW_scheduler.is_better(rmse, AdamW_scheduler.best):
no_improve_count = 0
torch.save(self.nn.state_dict(), self.model_checkpoint)
else:
Expand All @@ -619,17 +624,19 @@ def run(self):
if no_improve_count > self.max_nonimprove:
break

scheduler.step(rmse)
AdamW_scheduler.step(rmse)
SGD_scheduler.step(rmse)

if self.tensorboard is not None:
self.tensorboard.add_scalar('validation_rmse', rmse, scheduler.last_epoch)
self.tensorboard.add_scalar('best_validation_rmse', scheduler.best, scheduler.last_epoch)
self.tensorboard.add_scalar('learning_rate', learning_rate, scheduler.last_epoch)
self.tensorboard.add_scalar('no_improve_count_vs_epoch', no_improve_count, scheduler.last_epoch)
self.tensorboard.add_scalar('validation_rmse', rmse, AdamW_scheduler.last_epoch)
self.tensorboard.add_scalar('best_validation_rmse', AdamW_scheduler.best, AdamW_scheduler.last_epoch)
self.tensorboard.add_scalar('learning_rate', learning_rate, AdamW_scheduler.last_epoch)
self.tensorboard.add_scalar('no_improve_count_vs_epoch', no_improve_count, AdamW_scheduler.last_epoch)

for i, (batch_x, batch_y) in self.tqdm(
enumerate(self.training_set),
total=len(self.training_set),
desc='epoch {}'.format(scheduler.last_epoch)
desc='epoch {}'.format(AdamW_scheduler.last_epoch)
):

true_energies = batch_y['energies']
Expand All @@ -650,12 +657,12 @@ def run(self):

# write current batch loss to TensorBoard
if self.tensorboard is not None:
self.tensorboard.add_scalar('batch_loss', loss, scheduler.last_epoch * len(self.training_set) + i)
self.tensorboard.add_scalar('batch_loss', loss, AdamW_scheduler.last_epoch * len(self.training_set) + i)

# log elapsed time
elapsed = round(timeit.default_timer() - start, 2)
if self.tensorboard is not None:
self.tensorboard.add_scalar('time_vs_epoch', elapsed, scheduler.last_epoch)
self.tensorboard.add_scalar('time_vs_epoch', elapsed, AdamW_scheduler.last_epoch)


__all__ = ['Constants', 'load_sae', 'load_model', 'load_model_ensemble', 'Trainer']